actigraphy_cleaning
Jackie
5/31/2021; 6/11/21; 7/4/21; 8/4/21
This script wrangles actigraphy data for subsequent analyses # libraries
library(tidyverse)
library(naniar)
library(lubridate)
library(lmerTest)1 reading in databases
## # A tibble: 1 x 1
## start_date_max
## <date>
## 1 2021-04-25## # A tibble: 1 x 1
## start_date_max
## <date>
## 1 2020-08-20## # A tibble: 1 x 1
## start_date_max
## <date>
## 1 2020-08-142 cleaning databases
filter only for sleep intervals, and selecting only efficiency
options(scipen = 999)
# From Reed & Sacco
# DSE <- onset_latency + TST + WASO + TASAFA
# SE = (sleep_time/DSE)*100
d_vars <- c("data_start_date","onset_latency", "snooze_time", "efficiency", "duration", "sleep_time", "waso", "wake_time", "percent_wake", "number_of_wake_bouts", "avg_wake_b", "sleep_time", "number_of_sleep_bouts")
# d1
d1_clean <-
d1 %>%
rename(
ELS_ID = subject_id
) %>%
mutate(
ELS_ID = as.character(ELS_ID),
timepoint =
ifelse(
str_detect(ELS_ID, "T3"),
"T3",
NA
)
) %>%
relocate(timepoint, .after = ELS_ID) %>%
mutate(
ELS_ID = str_remove(ELS_ID, "-T3"),
ELS_ID = as.numeric(ELS_ID),
ELS_ID = factor(ELS_ID)
) %>%
filter(interval_type == "SLEEP") %>%
dplyr::select(ELS_ID,
timepoint,
data_start_date,
data_start_time,
interval_type,
interval_number,
start_date,
start_time,
end_date,
end_time,
duration,
onset_latency,
snooze_time,
efficiency,
waso,
wake_time,
percent_wake,
number_of_wake_bouts,
avg_wake_b,
sleep_time,
percent_sleep,
number_of_sleep_bouts,
avg_sleep_b,
inv_time_sw) %>%
filter(
!is.na(interval_number)
) %>%
filter(
inv_time_sw == 0
) %>%
mutate(
sleep_time_hrs = sleep_time/60
)
# d2
d2_clean <-
d2 %>%
rename(
ELS_ID = subject_id
) %>%
mutate(
ELS_ID = as.character(ELS_ID),
timepoint =
ifelse(
str_detect(ELS_ID, "T3"),
"T3",
NA
)
) %>%
relocate(timepoint, .after = ELS_ID) %>%
mutate(
ELS_ID = str_remove(ELS_ID, "-T3"),
ELS_ID = as.numeric(ELS_ID),
ELS_ID = factor(ELS_ID)
) %>%
filter(interval_type == "SLEEP") %>%
dplyr::select(ELS_ID,
timepoint,
data_start_date,
data_start_time,
interval_type,
interval_number,
start_date,
start_time,
end_date,
end_time,
duration,
onset_latency,
snooze_time,
efficiency,
waso,
wake_time,
percent_wake,
number_of_wake_bouts,
avg_wake_b,
sleep_time,
percent_sleep,
number_of_sleep_bouts,
avg_sleep_b,
inv_time_sw) %>%
filter(
!is.na(interval_number)
) %>%
filter(
inv_time_sw == 0
) %>%
mutate(
sleep_time_hrs = sleep_time/60
)
# d3
d3_clean <-
d3 %>%
rename(
ELS_ID = subject_id
) %>%
mutate(
ELS_ID = as.character(ELS_ID),
timepoint =
ifelse(
str_detect(ELS_ID, "T3"),
"T3",
NA
)
) %>%
relocate(timepoint, .after = ELS_ID) %>%
mutate(
ELS_ID = str_remove(ELS_ID, "-T3"),
ELS_ID = as.numeric(ELS_ID),
ELS_ID = factor(ELS_ID)
) %>%
filter(interval_type == "SLEEP") %>%
dplyr::select(ELS_ID,
timepoint,
data_start_date,
data_start_time,
interval_type,
interval_number,
start_date,
start_time,
end_date,
end_time,
duration,
onset_latency,
snooze_time,
efficiency,
waso,
wake_time,
percent_wake,
number_of_wake_bouts,
avg_wake_b,
sleep_time,
percent_sleep,
number_of_sleep_bouts,
avg_sleep_b,
inv_time_sw) %>%
filter(
!is.na(interval_number)
) %>%
filter(
inv_time_sw == 0
) %>%
mutate(
sleep_time_hrs = sleep_time/60
)
# d4
# in this database 2 ids were entered as each other by accident, so separating now by date administered correclty
### 109 was 6/23
### 171 was 7/8
d4_clean <-
d4 %>%
rename(
ELS_ID = subject_id
) %>%
mutate(
ELS_ID = as.character(ELS_ID),
timepoint =
ifelse(
str_detect(ELS_ID, "T3"),
"T3",
NA
)
) %>%
relocate(timepoint, .after = ELS_ID) %>%
mutate(
ELS_ID_chr =
ifelse(
str_detect(ELS_ID, "171-T3 and 109-T3") &
str_detect(data_start_date, "6/23/19"),
"109-T3",
NA
)
) %>%
relocate(ELS_ID_chr, .after = ELS_ID) %>%
mutate(
ELS_ID_chr =
ifelse(
str_detect(ELS_ID, "171-T3 and 109-T3") &
str_detect(data_start_date, "7/7/19"),
"171-T3",
ELS_ID_chr
)
) %>%
relocate(ELS_ID_chr, .after = ELS_ID) %>%
mutate(
ELS_ID_chr =
ifelse(is.na(ELS_ID_chr),
ELS_ID,
ELS_ID_chr)
) %>%
dplyr::select(-ELS_ID) %>%
rename(
ELS_ID = ELS_ID_chr
) %>%
mutate(
ELS_ID = str_remove(ELS_ID, "-T3"),
ELS_ID = as.numeric(ELS_ID),
ELS_ID = factor(ELS_ID)
) %>%
filter(interval_type == "SLEEP") %>%
dplyr::select(ELS_ID,
timepoint,
data_start_date,
data_start_time,
interval_type,
interval_number,
start_date,
start_time,
end_date,
end_time,
duration,
onset_latency,
snooze_time,
efficiency,
waso,
wake_time,
percent_wake,
number_of_wake_bouts,
avg_wake_b,
sleep_time,
percent_sleep,
number_of_sleep_bouts,
avg_sleep_b,
inv_time_sw) %>%
filter(
!is.na(interval_number)
) %>%
filter(
inv_time_sw == 0
) %>%
filter(
ELS_ID != "196" # using 196x in the cov df bc aligns with MW
) %>%
mutate(
sleep_time_hrs = sleep_time/60
)
# covid df
cov_clean <-
cov %>%
rename(
ELS_ID = subject_id
) %>%
mutate(
ELS_ID = as.character(ELS_ID),
timepoint =
ifelse(
str_detect(ELS_ID, "T3"),
"T3",
NA
),
timepoint =
ifelse(
str_detect(ELS_ID,"x-T3"),
"T3x",
timepoint
),
timepoint =
ifelse(
str_detect(ELS_ID, "T4"),
"T4",
timepoint
)
) %>%
relocate(timepoint, .after = ELS_ID) %>%
mutate(
ELS_ID = str_remove(ELS_ID, "-T3"),
ELS_ID = str_remove(ELS_ID, "-T4"),
ELS_ID = str_remove(ELS_ID, "x"),
ELS_ID = as.numeric(ELS_ID),
ELS_ID = factor(ELS_ID)
) %>%
filter(interval_type == "SLEEP") %>%
dplyr::select(ELS_ID,
timepoint,
data_start_date,
data_start_time,
interval_type,
interval_number,
start_date,
start_time,
end_date,
end_time,
duration,
onset_latency,
snooze_time,
efficiency,
waso,
wake_time,
percent_wake,
number_of_wake_bouts,
avg_wake_b,
sleep_time,
percent_sleep,
number_of_sleep_bouts,
avg_sleep_b,
inv_time_sw) %>%
filter(
!is.na(interval_number)
) %>%
filter(
inv_time_sw == 0
) %>%
mutate(
sleep_time_hrs = sleep_time/60
)
# participants administered actigraphy during COVID-19 from database 4
d4_cov_clean <-
d4_cov %>%
rename(
ELS_ID = subject_id
) %>%
mutate(
ELS_ID = as.character(ELS_ID),
timepoint =
ifelse(
str_detect(ELS_ID, "T3"),
"T3",
NA
),
timepoint =
ifelse(
str_detect(ELS_ID, "T4"),
"T4",
timepoint
)
) %>%
relocate(timepoint, .after = ELS_ID) %>%
mutate(
ELS_ID = str_remove(ELS_ID, "-T3"),
ELS_ID = str_remove(ELS_ID, "-T4"),
ELS_ID = str_remove(ELS_ID, "x"),
ELS_ID = as.numeric(ELS_ID),
ELS_ID = factor(ELS_ID)
) %>%
filter(interval_type == "SLEEP") %>%
dplyr::select(ELS_ID,
timepoint,
data_start_date,
data_start_time,
interval_type,
interval_number,
start_date,
start_time,
end_date,
end_time,
duration,
onset_latency,
snooze_time,
efficiency,
waso,
wake_time,
percent_wake,
number_of_wake_bouts,
avg_wake_b,
sleep_time,
percent_sleep,
number_of_sleep_bouts,
avg_sleep_b,
inv_time_sw) %>%
filter(
!is.na(interval_number)
) %>%
filter(
inv_time_sw == 0
) %>%
mutate(
sleep_time_hrs = sleep_time/60
) 3 extracting data start date for reduced dataframe
d1_clean_startdate_red <-
d1_clean %>%
dplyr::select(ELS_ID, data_start_date, timepoint) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date)
) %>%
dplyr::select(-data_start_date) %>%
mutate(ELS_ID = factor(ELS_ID))
d2_clean_startdate_red <-
d2_clean %>%
dplyr::select(ELS_ID, data_start_date, timepoint) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date)
) %>%
dplyr::select(-data_start_date) %>%
mutate(ELS_ID = factor(ELS_ID))
d3_clean_startdate_red <-
d3_clean %>%
dplyr::select(ELS_ID, data_start_date, timepoint) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date)
) %>%
dplyr::select(-data_start_date) %>%
mutate(ELS_ID = factor(ELS_ID))
d4_clean_startdate_red <-
d4_clean %>%
dplyr::select(ELS_ID, data_start_date, timepoint) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date)
) %>%
dplyr::select(-data_start_date) %>%
mutate(ELS_ID = factor(ELS_ID))
cov_clean_startdate_red <-
cov_clean %>%
dplyr::select(ELS_ID, data_start_date, timepoint) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date)
) %>%
dplyr::select(-data_start_date) %>%
mutate(ELS_ID = factor(ELS_ID))
d4_cov_clean_startdate_red <-
d4_cov_clean %>%
dplyr::select(ELS_ID, data_start_date, timepoint) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date)
) %>%
dplyr::select(-data_start_date) %>%
mutate(ELS_ID = factor(ELS_ID))4 extracting data start date for long datafame
d1_clean_startdate <-
d1_clean %>%
dplyr::select(ELS_ID, data_start_date, start_date) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date),
start_date = mdy(start_date),
actigraphy_day = day(start_date)
) %>%
dplyr::select(-data_start_date, -start_date) %>%
mutate(ELS_ID = factor(ELS_ID))
d2_clean_startdate <-
d2_clean %>%
dplyr::select(ELS_ID, data_start_date, start_date) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date),
start_date = mdy(start_date),
actigraphy_day = day(start_date)
) %>%
dplyr::select(-data_start_date, -start_date) %>%
mutate(ELS_ID = factor(ELS_ID))
d3_clean_startdate <-
d3_clean %>%
dplyr::select(ELS_ID, data_start_date, start_date) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date),
start_date = mdy(start_date),
actigraphy_day = day(start_date)
) %>%
dplyr::select(-data_start_date, -start_date) %>%
mutate(ELS_ID = factor(ELS_ID))
d4_clean_startdate <-
d4_clean %>%
dplyr::select(ELS_ID, data_start_date, start_date) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date),
start_date = mdy(start_date),
actigraphy_day = day(start_date)
) %>%
dplyr::select(-data_start_date, -start_date) %>%
mutate(ELS_ID = factor(ELS_ID))
cov_clean_startdate <-
cov_clean %>%
dplyr::select(ELS_ID, data_start_date, start_date) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date),
start_date = mdy(start_date),
actigraphy_day = day(start_date)
) %>%
dplyr::select(-data_start_date, -start_date) %>%
mutate(ELS_ID = factor(ELS_ID))
d4_cov_clean_startdate <-
d4_cov_clean %>%
dplyr::select(ELS_ID, data_start_date, start_date) %>%
unique() %>%
mutate(
actigraphy_start_date = mdy(data_start_date),
start_date = mdy(start_date),
actigraphy_day = day(start_date)
) %>%
dplyr::select(-data_start_date, -start_date) %>%
mutate(ELS_ID = factor(ELS_ID))4.1 merging long-form dataframes
# making start times the same object
d3_clean_time <-
d3_clean %>%
mutate(
start_time = as.character(start_time),
end_time = as.character(end_time)
)
d4_clean_time <-
d4_clean %>%
mutate(
start_time = as.character(start_time),
end_time = as.character(end_time)
)
cov_clean_time <-
cov_clean %>%
mutate(
start_time = as.character(start_time),
end_time = as.character(end_time)
)
d4_cov_clean_time <-
d4_cov_clean %>%
mutate(
start_time = as.character(start_time),
end_time = as.character(end_time)
)
actig_long <- bind_rows(d1_clean, d2_clean, d3_clean_time, d4_clean_time, cov_clean_time, d4_cov_clean_time)
actig_long_2wk <-
actig_long %>%
filter(
interval_number < 15
)actig_long_2wk <-
actig_long_2wk %>%
mutate(
triggerdate = mdy(start_date),
actstartdate = mdy(data_start_date)
)4.1.1 creating day and week variables
# creating day and week variables
actig_long_2wk_day <-
actig_long_2wk %>%
mutate(
day = day(triggerdate),
wday = wday(triggerdate, label = TRUE),
ELS_ID = factor(ELS_ID)
)
# day num
actig_long_2wk_dayorder <-
actig_long_2wk_day %>%
group_by(ELS_ID) %>%
mutate(
dayorder = order(triggerdate)
)4.1.2 wkday or wkend
actig_long_2wk_dayorder2 <-
actig_long_2wk_dayorder %>%
mutate(
week =
ifelse(
wday == "Sun" |
wday == "Sat",
"wkend",
"wkday"
)
)4.1.3 distribution of weekday vs. weekend and association with sleep eff and hrs
level_key <- c(wkday = "weekday", wkend = "weekend")
actig_long_2wk_dayorder2 <-
actig_long_2wk_dayorder2 %>%
mutate(
week = factor(week),
week = recode_factor(week,!!!level_key),
dayorder = as.numeric(dayorder),
sleep_time_hrs = as.numeric(sleep_time_hrs)
)
actig_long_2wk_dayorder2 %>%
ggplot(
aes(x = week, y = efficiency, fill = week)
) +
geom_violin(alpha=0.5, color= "black") +
geom_boxplot(width=0.1, color = "grey", alpha=0.5) +
scale_fill_manual(values = c("#FFA07A","#6B8E23")) +
theme_classic() +
labs(x = "Weekday or Weekend", y = "Sleep Effiency")
ggsave("wkdayvsweeknd_sleepeff.png", width = 7, height = 6)
summary(lmer(scale(efficiency) ~ scale(dayorder) +factor(week) + (1 + scale(dayorder)|ELS_ID), data = actig_long_2wk_dayorder2))## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula:
## scale(efficiency) ~ scale(dayorder) + factor(week) + (1 + scale(dayorder) |
## ELS_ID)
## Data: actig_long_2wk_dayorder2
##
## REML criterion at convergence: 4777.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -8.2256 -0.4075 0.1553 0.5907 2.3435
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## ELS_ID (Intercept) 0.35648 0.5971
## scale(dayorder) 0.01381 0.1175 0.05
## Residual 0.63570 0.7973
## Number of obs: 1860, groups: ELS_ID, 144
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.01518 0.05479 157.27870 0.277 0.7820
## scale(dayorder) -0.05090 0.02271 167.15509 -2.241 0.0263 *
## factor(week)weekend -0.06284 0.04134 1722.66288 -1.520 0.1287
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) scl(d)
## scal(dyrdr) 0.049
## fctr(wk)wkn -0.223 0.025actig_long_2wk_dayorder2 %>%
ggplot(
aes(x = week, y = sleep_time_hrs, fill = week)
) +
geom_violin(alpha=0.5, color= "black") +
geom_boxplot(width=0.1, color = "grey", alpha=0.5) +
scale_fill_manual(values = c("#FFA07A","#6B8E23")) +
theme_classic() +
labs(x = "Weekday or Weekend", y = "Objective Sleep Duration")
ggsave("wkdayvsweeknd_objsleephrs.png", width = 7, height = 6)
summary(lmer(scale(sleep_time_hrs) ~ scale(dayorder) + factor(week) + (1 + scale(dayorder)|ELS_ID), data = actig_long_2wk_dayorder2))## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: scale(sleep_time_hrs) ~ scale(dayorder) + factor(week) + (1 +
## scale(dayorder) | ELS_ID)
## Data: actig_long_2wk_dayorder2
##
## REML criterion at convergence: 5042.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.6205 -0.5237 0.0153 0.5649 4.1910
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## ELS_ID (Intercept) 0.210580 0.45889
## scale(dayorder) 0.008784 0.09372 -0.37
## Residual 0.773040 0.87923
## Number of obs: 1860, groups: ELS_ID, 144
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) -0.049484 0.045642 171.469961 -1.084 0.27981
## scale(dayorder) 0.008872 0.023289 174.180880 0.381 0.70369
## factor(week)weekend 0.140398 0.045342 1736.431852 3.096 0.00199 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) scl(d)
## scal(dyrdr) -0.082
## fctr(wk)wkn -0.294 0.027sleep_week <-
actig_long_2wk_dayorder2 %>%
group_by(ELS_ID, week) %>%
summarize(
n = n()
)## `summarise()` has grouped output by 'ELS_ID'. You can override using the `.groups` argument.actig_long_2wk_dayorder2_spread <-
actig_long_2wk_dayorder2 %>%
spread(week, efficiency)
actig_long_2wk_dayorder2 %>%
ggplot(
aes(
x = dayorder,
y = efficiency,
color = week
)
) +
geom_line(aes(group = ELS_ID, color = week), alpha = .3) +
geom_smooth(method = "loess", se=FALSE, size = 2) +
theme_classic() +
scale_x_continuous(
name = "Day",
limits = c(1, 14),
breaks = seq(1, 14, 1)
) +
scale_y_continuous(
name = "Sleep Efficiency",
limits = c(1, 100),
breaks = seq(1, 100, 10)
) +
scale_color_manual(values = c("#FFA07A","#6B8E23"))## `geom_smooth()` using formula 'y ~ x'## Warning: Removed 33 rows containing non-finite values (stat_smooth).## Warning: Removed 33 row(s) containing missing values (geom_path).
ggsave("sleepeff_byDay_byWkdayWkend.png", width = 7, height = 6)## `geom_smooth()` using formula 'y ~ x'## Warning: Removed 33 rows containing non-finite values (stat_smooth).
## Warning: Removed 33 row(s) containing missing values (geom_path).actig_long_2wk_dayorder2 %>%
ggplot(
aes(
x = dayorder,
y = sleep_time_hrs,
color = week
)
) +
geom_line(aes(group = ELS_ID, color = week), alpha = .3) +
geom_smooth(method = "loess", se=FALSE, size = 2) +
theme_classic() +
scale_x_continuous(
name = "Day",
limits = c(1, 14),
breaks = seq(1, 14, 1)
) +
scale_y_continuous(
name = "Sleep Hours",
limits = c(0, 13),
breaks = seq(0, 13, 1)
) +
scale_color_manual(values = c("#FFA07A","#6B8E23"))## `geom_smooth()` using formula 'y ~ x'## Warning: Removed 33 rows containing non-finite values (stat_smooth).
## Warning: Removed 33 row(s) containing missing values (geom_path).
ggsave("objsleephrs_byDay_byWkdayWkend.png", width = 7, height = 6)## `geom_smooth()` using formula 'y ~ x'## Warning: Removed 33 rows containing non-finite values (stat_smooth).
## Warning: Removed 33 row(s) containing missing values (geom_path).4.2 merging reduced dataframes
5 reducing dataframes
taking the average of efficiency over the span of the 14 days to get a more “representative” efficiency
# d1
d1_reduced <-
d1_clean %>%
mutate(
ELS_ID = as.factor(ELS_ID)
) %>%
group_by(ELS_ID) %>%
summarize(
eff_perc_mean =
mean(efficiency, na.rm = TRUE),
eff_perc_sd =
sd(efficiency, na.rm = TRUE),
onset_lat_mean =
mean(onset_latency, na.rm = TRUE),
onset_lat_sd =
sd(onset_latency, na.rm = TRUE),
snooze_mean =
mean(snooze_time, na.rm = TRUE),
snooze_sd =
sd(snooze_time, na.rm = TRUE),
sleep_dur_mean =
mean(sleep_time, na.rm = TRUE),
sleep_dur_sd =
sd(sleep_time, na.rm = TRUE),
sleep_time_hrs_mean =
mean(sleep_time_hrs, na.rm = TRUE),
sleep_time_hrs_sd =
sd(sleep_time_hrs, na.rm = TRUE)
)
d1_reduced <-
left_join(
d1_clean_startdate_red,
d1_reduced,
by = "ELS_ID"
)
# d2
d2_reduced <-
d2_clean %>%
mutate(
ELS_ID = as.factor(ELS_ID)
) %>%
group_by(ELS_ID) %>%
summarize(
eff_perc_mean =
mean(efficiency, na.rm = TRUE),
eff_perc_sd =
sd(efficiency, na.rm = TRUE),
onset_lat_mean =
mean(onset_latency, na.rm = TRUE),
onset_lat_sd =
sd(onset_latency, na.rm = TRUE),
snooze_mean =
mean(snooze_time, na.rm = TRUE),
snooze_sd =
sd(snooze_time, na.rm = TRUE),
sleep_dur_mean =
mean(sleep_time, na.rm = TRUE),
sleep_dur_sd =
sd(sleep_time, na.rm = TRUE),
sleep_time_hrs_mean =
mean(sleep_time_hrs, na.rm = TRUE),
sleep_time_hrs_sd =
sd(sleep_time_hrs, na.rm = TRUE)
)
d2_reduced <-
left_join(
d2_clean_startdate_red,
d2_reduced,
by = "ELS_ID"
)
# d3
d3_reduced <-
d3_clean %>%
mutate(
ELS_ID = as.factor(ELS_ID)
) %>%
group_by(ELS_ID) %>%
summarize(
eff_perc_mean =
mean(efficiency, na.rm = TRUE),
eff_perc_sd =
sd(efficiency, na.rm = TRUE),
onset_lat_mean =
mean(onset_latency, na.rm = TRUE),
onset_lat_sd =
sd(onset_latency, na.rm = TRUE),
snooze_mean =
mean(snooze_time, na.rm = TRUE),
snooze_sd =
sd(snooze_time, na.rm = TRUE),
sleep_dur_mean =
mean(sleep_time, na.rm = TRUE),
sleep_dur_sd =
sd(sleep_time, na.rm = TRUE),
sleep_time_hrs_mean =
mean(sleep_time_hrs, na.rm = TRUE),
sleep_time_hrs_sd =
sd(sleep_time_hrs, na.rm = TRUE)
)
d3_reduced <-
left_join(
d3_clean_startdate_red,
d3_reduced,
by = "ELS_ID"
)
# d4
d4_reduced <-
d4_clean %>%
mutate(
ELS_ID = as.factor(ELS_ID)
) %>%
group_by(ELS_ID) %>%
summarize(
eff_perc_mean =
mean(efficiency, na.rm = TRUE),
eff_perc_sd =
sd(efficiency, na.rm = TRUE),
onset_lat_mean =
mean(onset_latency, na.rm = TRUE),
onset_lat_sd =
sd(onset_latency, na.rm = TRUE),
snooze_mean =
mean(snooze_time, na.rm = TRUE),
snooze_sd =
sd(snooze_time, na.rm = TRUE),
sleep_dur_mean =
mean(sleep_time, na.rm = TRUE),
sleep_dur_sd =
sd(sleep_time, na.rm = TRUE),
sleep_time_hrs_mean =
mean(sleep_time_hrs, na.rm = TRUE),
sleep_time_hrs_sd =
sd(sleep_time_hrs, na.rm = TRUE)
)
d4_reduced <-
left_join(
d4_clean_startdate_red,
d4_reduced,
by = "ELS_ID"
)
# cov
cov_reduced <-
cov_clean %>%
mutate(
ELS_ID = as.factor(ELS_ID)
) %>%
group_by(ELS_ID) %>%
summarize(
eff_perc_mean =
mean(efficiency, na.rm = TRUE),
eff_perc_sd =
sd(efficiency, na.rm = TRUE),
onset_lat_mean =
mean(onset_latency, na.rm = TRUE),
onset_lat_sd =
sd(onset_latency, na.rm = TRUE),
snooze_mean =
mean(snooze_time, na.rm = TRUE),
snooze_sd =
sd(snooze_time, na.rm = TRUE),
sleep_dur_mean =
mean(sleep_time, na.rm = TRUE),
sleep_dur_sd =
sd(sleep_time, na.rm = TRUE),
sleep_time_hrs_mean =
mean(sleep_time_hrs, na.rm = TRUE),
sleep_time_hrs_sd =
sd(sleep_time_hrs, na.rm = TRUE)
)
cov_reduced <-
left_join(
cov_clean_startdate_red,
cov_reduced,
by = "ELS_ID"
)
# d4 cov
d4_cov_reduced <-
d4_cov_clean %>%
mutate(
ELS_ID = as.factor(ELS_ID)
) %>%
group_by(ELS_ID) %>%
summarize(
eff_perc_mean =
mean(efficiency, na.rm = TRUE),
eff_perc_sd =
sd(efficiency, na.rm = TRUE),
onset_lat_mean =
mean(onset_latency, na.rm = TRUE),
onset_lat_sd =
sd(onset_latency, na.rm = TRUE),
snooze_mean =
mean(snooze_time, na.rm = TRUE),
snooze_sd =
sd(snooze_time, na.rm = TRUE),
sleep_dur_mean =
mean(sleep_time, na.rm = TRUE),
sleep_dur_sd =
sd(sleep_time, na.rm = TRUE),
sleep_time_hrs_mean =
mean(sleep_time_hrs, na.rm = TRUE),
sleep_time_hrs_sd =
sd(sleep_time_hrs, na.rm = TRUE)
)
d4_cov_reduced <-
left_join(
d4_cov_clean_startdate_red,
d4_cov_reduced,
by = "ELS_ID"
)actig_merged <- bind_rows(d1_reduced, d2_reduced, d3_reduced, d4_reduced, cov_reduced, d4_cov_reduced)6 Data Viz
6.1 Day to Day sleep efficiency (software based)
actig_long_2wk %>%
ggplot(
aes(
interval_number,
efficiency,
group = ELS_ID
)
) +
geom_line(
alpha = 1/2
) +
facet_wrap(~ELS_ID) +
theme_minimal() +
theme(
strip.text.x = element_blank()
) +
labs(
x = "Day",
y = "Sleep Efficiency"
) +
scale_y_continuous(limits = c(0,100))
ggsave("daily_act_sleep_eff_traj.png", width = 7, height = 7, dpi = 400)6.2 Day to Day sleep dur
actig_long_2wk %>%
ggplot(
aes(
interval_number,
sleep_time_hrs,
group = ELS_ID
)
) +
geom_line(
alpha = 1/2
) +
facet_wrap(~ELS_ID) +
theme_minimal() +
theme(
strip.text.x = element_blank()
) +
labs(
x = "Day",
y = "Sleep Time (hrs)"
) 
ggsave("daily_act_sleep_time_traj.png", width = 7, height = 7, dpi = 400)7 sleep instability?
computing the rmssd as per Koval et al., 2013; Jahng et al., 2008 The RMSSD is measured as the square root of the average of the squared differences between affect at measurement i and i + 1 (Schoevers et al., 2020)
sleep_lag_df <-
actig_long_2wk %>%
dplyr::select(
ELS_ID,interval_number, efficiency, sleep_time, sleep_time_hrs
) %>%
group_by(ELS_ID) %>%
summarize(
# step 1: computing successive difference
sleep_hrs_succ_diff = sleep_time_hrs - lag(sleep_time_hrs),
# step 2: computing square of each diff
sleep_hrs_sq_succ_diff = sleep_hrs_succ_diff^2
)## `summarise()` has grouped output by 'ELS_ID'. You can override using the `.groups` argument.sleep_mssd_df <-
sleep_lag_df %>%
group_by(ELS_ID) %>%
summarize(
# step 3: averaging the squared differences
mean_sq_sleep_hrs_succ_diff = mean(sleep_hrs_sq_succ_diff, na.rm = TRUE),
# step 4: computing square root of the average
rmssd_sleep_hrs = sqrt(mean_sq_sleep_hrs_succ_diff)
) %>%
mutate(ELS_ID = factor(ELS_ID)) 7.0.0.1 joining dataframes
actig_sleep <-
left_join(
actig_merged,
sleep_mssd_df,
by = "ELS_ID"
) %>%
relocate(actigraphy_start_date, .after = ELS_ID)7.1 summary stats
actig_sleep_summary <-
actig_sleep %>%
summarize(
n = n(),
sleep_eff_mean = mean(eff_perc_mean),
sleep_eff_sd = sd(eff_perc_mean),
sleep_time_mean = mean(sleep_dur_mean),
sleep_time_sd = sd(sleep_dur_mean),
sleep_time_hrs_mean = mean(sleep_time_hrs_mean),
sleep_time_hrs_sd = sd(sleep_time_hrs_mean)
)
write_csv(actig_sleep_summary, "actig_sleep_summary_stats.csv") 8 Data Viz
8.1 distributions of sleep eff
# mean
actig_sleep %>%
ggplot(
aes(x = eff_perc_mean)
) +
geom_histogram(alpha = .5, color="black") +
theme_classic() ## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# sd
actig_sleep %>%
ggplot(
aes(x = eff_perc_sd)
) +
geom_histogram(alpha = .5, color = "black") +
theme_classic() ## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
8.2 distributions of sleep time
# mean
actig_sleep %>%
ggplot(
aes(x = sleep_time_hrs_mean)
) +
geom_histogram(alpha = .5, color="black") +
theme_classic() ## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# rmssd
actig_sleep %>%
ggplot(
aes(x = rmssd_sleep_hrs)
) +
geom_histogram(alpha = .5, color="black") +
theme_classic() ## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
8.2.1 writing out reduced and long form actigraphy dataframe
write_csv(actig_sleep, "~/Box/Mooddata_Coordinating/1_Lab_Coordinating/Users/JackieSchwartz/Dissertation/0_MW_Act_Demo_Descriptives/actig_merged.csv")
write_csv(actig_long_2wk, "~/Box/Mooddata_Coordinating/1_Lab_Coordinating/Users/JackieSchwartz/Dissertation/0_MW_Act_Demo_Descriptives/actig_long.csv")